Jeff Ding introduced friction stir welding (FSW) to NASA in 1995. He currently holds 6 U.S. patents for FSW, including one for an automatic retractable pin tool that solves the troublesome “keyhole” problem. He is also credited with inventing two new solid state welding processes called thermal stir welding (TSW) and ultrasonic stir welding (USW). Ding was Marshall Space Flight Center’s Inventor of the Year in 2000, was awarded the Medal for Exceptional Technology Achievement in 2003, and recently received the 2009 Federal laboratory Consortium Award for Excellence in Technology Transfer.
NASA Tech Briefs: Jeff, you are considered to be NASA’s foremost expert in the field of friction stir welding (FSW). What is friction stir welding, and how does it work?
Jeff Ding: Friction stir welding is a solid state weld process, meaning you do not melt the material when you weld. The material is heated into a plastic state – a temperature between solidus and liquidus – using a rotating FSW pin tool. The pin tool has a larger diameter shoulder (approximately three times material thickness) from which a smaller diameter threaded pin protrudes. The threaded pin diameter is approximately equal to the thickness of material being welded and the pin length is approximately .025-in shorter than the thickness of material being welded. Essentially, if you’re welding aluminum, you slowly plunge the rotating, protruding pin into the weld joint until the shoulder comes in contact with the surface of the weld piece. After dwelling for a short period, the shoulder’s frictional energy creates enough heat to allow the aluminum to become soft, pliable, and “mushy” (plastic). At this time, the rotating pin tool (shoulder and pin) traverses the weld joint allowing the threaded pin inside the weld joint material to “stir” the abutting joint material. For .300-in thick aluminum, the pin tool rotates approximately 250 to 300 rpm and traverses the weld joint approximately 6 inches per minute.
NTB: I see. Although you’re credited with introducing friction stir welding to NASA in 1995, the technology was actually invented in 1991 by a man named Wayne Thomas and his team of researchers at The Welding Institute (TWI) in Great Britain. How did you get involved with friction stir welding?
Ding: Back in 1995 I was working in the space shuttle main engine (SSME) chief engineer’s office. I had been there for seven years. It was kind of ironic because I hired into the welding group in 1986 here at Marshall and spent my first two years completing the professional internship program (PIP), a program all new engineers fresh out of college must complete. During those two years, several three month rotations are spent in other MFSC organizations. I rotated to the SSME Office, where, after completing my PIP program, I was offered a permanent job. I guess they liked my welding background; at the time there were over 10,000 welds in a space shuttle main engine. So, I worked there for seven years, and through the years the welding people periodically asked me to come back to the welding organization because they were busy and needed help. So after seven years I thought, “Yeah, I’m going to head off back to the welding group.”
I believe it was in October of that year (1995) that I heard about this new welding process. To be able to do long, linear welds without melting the material seemed very intriguing, because when you don’t melt the material you get much higher properties out of the weld joint as compared to those weld processes where you melt the material. By the way, the processes where you melt material are referred to as fusion weld processes. Metal inert gas (MIG), tungsten inert gas (TIG), variable polarity plasma arc (VPPA), electron beam (EB) are all fusion weld processes, whereas friction stir welding is solid state.
I heard about this and the first thing I did was I called Edison Weld Institute (EWI) in Columbus, Ohio. At the time, EWI was a sister organization to The Welding Institute (TWI) in Cambridge, UK. By belonging to EWI, we could call TWI over in England for consultation. They were sister organizations, so being a member of one allowed you to use the services of the other. So I called Edison and they really didn’t know much about FSW. There were very few people back in 1995 who knew about this process. They said, “You’re going to have to call TWI.” So I called TWI and spoke with Dave Nicholas, who gave me a brief education, and I thought, this is great stuff!
So, I came back to the weld group in November 1995 and I pursued friction stir welding. The first thing I needed was a friction stir welding machine, but there were none available on the market since the technology was so new. What people did back in the 1990’s was use a CNC milling machine with “beefed up” bearings so that the machine could sustain the high forces associated with the technology. For the .300-in thick aluminum I spoke about earlier, you’re pushing down with probably 5,000 or 6,000 psi, so you need that structural, robust tooling – the anvil – to support the workpiece and react the load. Not many machines can sustain forces like that during operation.
So, I needed a milling machine, and I’d just left my position in the SSME Chief Engineer’s Office, where, for seven years, I interfaced with the SSME prime contractor, Rocketdyne, located in Canoga Park, CA. I knew Rocketdyne, was going through this big factory equipment divestment program because much of the SSME engine hardware had already been made and delivered to NASA. I visited Rocketdyne and found the perfect machine. It was a 14-ton Kearney & Trecker CNC horizontal boring mill, and it belonged to Rocketdyne. I offered to buy it, but I guess buying a contractor’s piece of equipment would’ve taken years in red tape, so I said, “Well, why don’t you guys just give that to me?” I was kind of being facetious, but they said, “Well, Mr. Ding, we’ll look into that.” The next day I got a call and they said they’d noticed that the book value on the thing was only worth $500 to them, so they said, “It’s yours.”
NTB: So it was easier for them to give it to you than it would’ve been for them to sell it to you?
Ding: Yes, much easier. After I knew the milling machine was mine, I visited my previous boss, Otto Goetz, SSME Chief Engineer. I knew there was a set of instrumented ducts coming to MSFC from Rocketdyne for our SSME testbed engine. I asked Otto if it would be possible to ship my machine along with the instrumented ducts. He said, “Well sure. We’ll do that.” So a rigger disassembled, boxed up, and put this big 14-ton Kearney & Trecker horizontal boring mill on a flatbed truck for me, and for $8,000 I had that thing sitting on the doorstep. The local Rocketdyne guys hooked it up for me, and I was in business and made my first weld, probably I think, in November 1996.
NTB: Being the first, that made you “the expert,” right?
Ding: Well, there weren’t a whole heck of a lot of people looking at the process. Back then this was just a lab curiosity, nothing more. It showed promise, and what I did was I participated in TWI’s Group Sponsored Projects. There were three Phases. The “Group” consisted of about 20 companies, worldwide. We each contributed around $50,000 (US) for the Phase I study that took FSW from a lab curiosity into a very low technical readiness level.
Phases II and III follow-on efforts continued for several more years. By the end of Phase III this [technology] was mature enough that people could take all of the data that TWI generated and pursue the technology for their own applications. By participating in those group-sponsored projects, NASA has a significantly reduced licensing fee for the life of that patent.
NTB: Why is friction stir welding preferred over more conventional fusion types of welding in certain applications?
Ding: Well, the number one reason is the fact that you aren’t melting the material, so you get much higher weld mechanical properties from the weld joint as compared to the fusion weld processes. That’s the primary reason.
Another reason is that it’s a highly automated process and very repeatable. If there were ever a weld process where you just push the button and let her go, this is it. There are four process variables - the pin rotational speed (RPM), travel rate, the force that the shoulder is pushing into the part, and the position of the shoulder below the surface of the weld piece. They are all highly computer controlled. Once you develop a parameter window and find the optimal parameters, you will perform the same weld over and over. The reliability and repeatability of manufacturing processes is very important when fabricating man-rated space hardware.
NTB: I assume you don’t have to heat treat it afterwards?
Ding: No. In some instances you might want to heat treat the whole part, at which point you could heat treat the weld itself. But for our applications here, for the aluminum that we’re welding for Ares I and for the external tank that we’ve used this for, it’s just as-welded. So we get better properties out of the weld joint. The design people and structural people love it because it helps them with their margins of safety. They appreciate that.
NTB: Are there certain materials or types of parts that are better suited to being joined by friction stir welding than other processes, or will the technique work just as well on anything?
Ding: Well, it was originally designed for aluminum, and it works very well with aluminum. It works well with joining the same alloy of aluminum, and it works well with joining dissimilar alloys of aluminum – say a 2195 to a 2219.
It works very well for joining those alloys considered unweldable, like your 7000 series aluminum alloys. You really can’t fusion weld them – the resulting weld properties are terrible. The aircraft industry uses a lot of 7000 series aluminum and, until recently, the material has always been riveted, not welded. The new Eclipse jet, however, is welding 7000 series aluminum. But with friction stir welding you get acceptable properties. And you can weld 7000 series to 2000 series, so it works well for aluminum.
High-melting-temperature alloys, such as steel, Inconels, and titanium, present a special challenge. Welding temperatures can be as high as 2000 degrees F, and when welding, the pin tool glows bright red. The elevated temperatures in which the pin tool operates tend to wear the pin tool excessively, and, in thicker material, it’s difficult to complete welds greater than a few feet in length. The pin tool materials are different than those used to weld aluminum alloys. H-13 tool steel and MP-159 are pretty much standard materials to make pin tools for the welding of aluminum. For these other materials (steel, Inconels, titanium, etc.), lanthinated tungsten, tungsten rhenium, and PCBN are used for pin tool materials.
NTB: Speaking of the pin, one of the major problems with friction stir welding that you’ve been credited with solving was the exit hole — also known as the keyhole — that was left in the part when the welding tool was withdrawn. Tell us about that problem and how you solved it.
Ding: Well, early on, back in 1996, when I was making my first welds, that was the first thing that kind of hit me. I thought, well, we’ve got to do something with this hole. So, we thought we needed this pin tool where the pin moves independently in and out of the shoulder. And I remember it was back in, oh, I think maybe it was 1997…it could’ve been 1996…President Clinton shut the Center down. He just closed the government down because he was battling Congress for his budget. Here at Marshall Space Flight Center, as well as all the NASA facilities, they chained the doors shut. So, while MSFC was closed, one of the engineers, Peter Oelgoetz, who worked for Rocketdyne and who was also my partner in crime in developing this process, had to report to an offsite building here in Huntsville. It was during that period – the week or so we were shut down – we put our heads together and decided that this was the design of how we wanted this thing to work. By the way, Pete is on the patent with me – I can’t take full credit because it was the two of us. But I can say, if you asked “Who got the idea to do it?” I said “Let’s do this.” So, we submitted the patent application for the “Auto-Adjustable Pin Tool for Friction Stir Welding.” More simply, it’s called the retractable pin tool, or RPT.
But the two of us put our heads together and figured out how to do it. We did a manual hand crank job, just a hand crank [that moved] the little pin up into the shoulder. We didn’t know if this concept was going to work and we didn’t want to spend a whole lot of money, so we just operated it by hand. It worked great, so we came back with modifications to automate it. The second design iteration was automated but not robust enough, so we came out with the third design change. We made two retractable pin tools, and they were very robust. One was made for my Kearney & Trecker horizontal boring mill and one was made for our vertical weld tool in building 4705.
So now we had a robust mechanical devise that could move the pin in and out of the shoulder under the axial forces associated with the FSW process. One of the benefits of the device is that it provided the mechanical means to weld material that tapered from one thickness to another. Variable thickness, or, tapered weld joints are designed into the Shuttle External Tank where the boosters and orbiter attach. For example, in the liquid hydrogen barrel section, weld joints taper in thickness from .320-in, to .550-in, to 1.00-in, to .650-in, to 1.00-in, to .550-in and finally back to .320-in. So, the pin must be able to automatically extend and retract in and out of the shoulder as it traverses the different weld thicknesses. The tip of the pin must always be positioned within about .030-in from the back side of the weld joint material.
It also allowed us to close out the keyhole in circumferential welds. Once the pin tool traverses the full 360 degree weld joint, it goes past the starting point and slowly retracts the pin into the shoulder, thus, closing out the keyhole.
A third benefit may be the most important in that it provides the mechanical means to operate the self-reacting friction stir weld (SR-FSW) pin tool technology. The SR-FSW is comprised of two shoulders – one rotates on the front side of the weld joint material and the other on the backside. They rotate together in tandem at the same RPM inducing frictional energy into the part from the front and backside surfaces. The conventional FSW pin tool has only one shoulder that rotates and creates frictional energy on the front side of the weld material. The conventional, single shouldered FSW process requires a robust, expensive anvil to “push” against while traversing the weld. The “pushing” forces exerted by the shoulder can approach 14,000 PSI when welding 1.00-in thick aluminum. One can imagine the structural hardware (anvil) required to react against such a force. The SR-FSW technique, on the other hand, does not require the reactive tooling, because there are no “pushing” forces. Instead, the two shoulders, (with the weld joint material sandwiched between them), “squeeze” together, thus, creating an equivalent “pinch” force equal to the “push” force of the conventional FSW pin tool. The Auto-Adjustable Pin Tool for Friction Stir Welding provides the mechanical means to retract one of the shoulders to make it self-react against the other, thus, creating the “pinch force”.
NTB: Since the advent of friction stir welding, two similar stir welding technologies – thermal stir welding and ultrasonic stir welding – have been developed at NASA. How do these two technologies differ from friction stir welding?
Ding: Thermal stir welding was an idea I came up with while developing FSW. In conventional FSW, the shoulder provides much of the frictional energy (heat) as well as the compressive, or forging force. The pin, which is attached to the shoulder, “stirs” the weld joint material together. These three FSW process elements – stirring, forging and heating – work in tandem at a desired RPM and cannot be decoupled from each other. The thermal stir welding (TSW) process de-couples the heating, stirring, and forging elements and allows for individual control of each. This allows for greater process control. I’ll give an example of what I think the benefit is of TSW.
When using conventional FSW to weld .500-in thick commercially pure (CP) titanium, the FSW pin tool must rotate between 700 – 900 RPM to generate the frictional energy required to plasticize the material. With thermal stir welding, I first heat the part with a specially designed induction coil; it heats very quickly through the thickness. Once you’re up to some temperature – it could be 1400 degrees, it could be 1600 degrees – whatever temperature it is that you want to stir, you move the part into your stir rod and all that stir rod does is it stirs the material together just like the little pin on a friction stir welding pin tool. The stir rod protrudes through the middle of two containment plates that contain the material as it is being stirred. Containment plates are stationary – they do not rotate. The containment plates also supply the compressive load to the stirred material for microstructure consolidation. The force they compress with is also controlled independently. Yes, I’m getting what we call adiabatic heating from the friction given off by the stir rod, but the primary source of heating is the induction coil. I have done welds rotating at only 200 RPM in ½-inch thick titanium.
Now, what’s the benefit of independent control of heating, stirring and forging pressures? It certainly increases life of the stir rod. Since the material is already at temperature when the stirring begins, the stir rod is primarily just a mixing tool that moves plastic material within the weld zone. In FSW, the shoulder/pin assembly must provide both heat and stir functions. This reduces life of the shoulder/pin assembly at the temperatures required to join CP Ti. Other benefits of independent control I am not sure of yet. TSW is very new. I’m just now looking at the .500-in thick CP Ti weld microstructures to compare the microstructures that were stirred using 100 RPM, 200 RPM, and 300 RPM, to see if I can see any differences. What I suspect, is that the strain rate being induced into the microstructure with the TSW process is much less than using FSW. I have data that shows the FSW pin tool rotational speed to heat and stir .500-in thick CP Ti must be between 700-900 RPM. Since commercially pure titanium has no alloying elements, the microstructure – the grains – are free to grow very large very quickly. The high strain rate induced into the microstructure creates tears and results in wormholes. With thermal stir welding, you can rotate very slowly and put a lot less strain into the microstructure. I believe that’s one reason why we’re getting good welds on .500-in thick titanium. I just got done completing an eight (8) foot long weld in .500-in thick CP Ti. Visually, it looks great! Radiography will tell the real story. By the way, the work I’m doing with the .500-in thick CP Ti is for the Office of Naval Research (ONR) – not NASA programs. The technology used to support the Constellation Program must be at a high TRL. TSW isn’t there yet, but I’m getting there. I am going to do a study to compare welding Haynes 230 with FSW and TSW. This will be my first work using TSW to support the Constellation Program.
Ding: So right now, TSW development is being done for ONR. ONR funds the TSW work through an SBIR with Keystone Synergistic Enterprises, Inc., Port St. Lucie, Fl. Keystone, in turn, funds MSFC through a reimbursable Space Act Agreement. I develop the process, provide Keystone with all the data that is confidential to them, and then they present it to the Navy. The Navy is very interested in this process, because the Navy is looking at this low-cost titanium product that’s processed a different way than your usual titanium. Instead of costing $60 a pound, it’ll cost $5 a pound. The reduced cost comes from the processing of the titanium. When the titanium is processed, there are a lot of tramp elements left in the metal such as chlorine. This presents problems when welding the low cost titanium. When it is welded with fusion weld processes, like TIG, MIG, or electron beam, a lot of oxides form, resulting in inferior weld properties. It has to be welded without melting, meaning, a solid state process such as FSW or TSW must be used. So far, ONR is quite impressed with TSW.
NTB: What is ultrasonic stir welding?
Ding: First of all I’m happy to say that the U.S. patent was allowed for the ultrasonic stir welding (USW) process just a few weeks ago. USW is also a solid state welding process. I heat the weld piece using high powered ultrasonic energy. Ultrasonic stir welding is an idea I got after looking at video of ultrasonic assisted drilling. The video shows the effects of drilling through, say, a half-inch thick steel plate with a quarter-inch drill bit, with and without ultrasonic energy applied. Without ultrasonic energy, the drill takes considerable time and force to drill through the steel plate. A load cell records the amount of force it takes to push the drill clear through the plate and is represented in a strip chart. During the drilling with no ultrasonics, the needle on the strip chart goes very high on the scale and then drops to zero when the drill pops through the steel plate. When the ultrasonic energy is applied, the drill cuts through the steel plate very quickly with significantly less force. The strip chart needle barely rises from the bottom of the scale. Without ultrasonic energy, the typical metal chips fly off the drill bit while drilling. When the ultrasonic energy is turned on, one long metallic “apple peel” is discharged, or, peels away, from the drilling process – no chips. This is because the ultrasonic energy plasticizes the steel at the interface between the drill tip and the workpiece. It’s really pretty amazing! So this video showed the heating of the metal by the drill bit and the reduction of forces when drilling. And what is USW? It’s primarily a very, very small drill bit (to stir the plastic material) with a non rotating containment plate to contain the plastic material.
Last summer I set up a little high-powered ultrasonic test bed to generate data. The primary data showed that I can heat metals into a plastic temperature state with ultrasonic energy and I can significantly reduce plunging forces. So ultrasonic stir welding, I believe, will be a way that we can take a solid state weld process and integrate it with an off-the-shelf robot for welding. Right now a huge robust robot is required to absorb the loads for friction stir welding, but with ultrasonics, I think, an off-the-shelf robot will be able to perform USW
NTB: What types of commercial applications do you envision for these technologies, and how long do you think it will take to commercialize them?
Ding: Well, of course, I look at NASA’s interests. We’re going to the moon and hopefully Mars. One objective of going to the moon is that we have to learn to live off the land and show that we can make parts, and make repairs. With long duration space travel, like going to Mars, you can’t take vehicles full of spare parts, so you have to be able to make things on the fly. Welding is a big part of making things. If an engine part fails, for example, there must be capability to repair it or make a new part. Welding is a big part of manufacturing in space. We know that, based on some previous work we did back in the 90s with the Russians with an electron beam welding experiment that we conducted here at Marshall, we learned that electron beam welding and fusion weld processes are too dangerous for the astronauts. We don’t like that, so we have to go to solid state welding. I believe that ultrasonics will allow us to have a stir weld process that will work using minimal forces. And anything we can do to reduce manual forces in the weightlessness of space would be a benefit.
Another application would be for simple robotic control, for different fields of use. That’s another area for the ultrasonics. Right now, with thermal stir welding, I believe it could be a significant advancement in welding technology, especially for your high-melting-temperature alloys that are difficult to weld with friction stir welding.